Topics as Speech Acts – An Analysis of Conditionals
Transcription
Topics as Speech Acts – An Analysis of Conditionals
Topics as Speech Acts – An Analysis of Conditionals Christian Ebert (Universität Bielefeld) Cornelia Endriss (Universität Osnabrück) Stefan Hinterwimmer (Humboldt-Universität zu Berlin) Overview introduction frame setting setting –– conclusion conclusion introduction –– topics topics as as speech speech acts acts –– ICs ICs as as aboutness aboutness topics topics –– BCs BCs as as frame We show syntactic similarities Indicative Conditionals of two types of conditionals and two types of left German Left Dislocation dislocation constructions in German, Aboutness Topicality which mark two types of topicality. Biscuit Conditionals Hanging Topic Left Dislocation Frame Setting On basis of these similarities we argue that (the antecedents of) Indicative Conditionals are Aboutness Topics Biscuit Conditionals are Frame Setting Topics ► We provide a uniform derivation of the semantic and pragmatic contributions of the two conditional forms by extending the aboutness topicality approach of Endriss (to appear) to frame setting topics and combining it with the analysis of indicative conditionals of Schlenker (2004) WCCFL 27, 16-18 May 2008, UCLA 2 Indicative Conditionals vs. Biscuit Conditionals introduction frame setting setting –– conclusion conclusion introduction –– topics topics as as speech speech acts acts –– ICs ICs as as aboutness aboutness topics topics –– BCs BCs as as frame Indicative Conditional (IC): (1) If Peter went shopping, then there is pizza in the fridge. IC BC GLD HTLD Aboutness Frame Setting Truth of the consequent depends on the truth of the antecedent. Biscuit Conditional (BC): (2) If you are hungry, (*then) there is pizza in the fridge. Truth of the consequent is independent of the truth of the antecedent. Characteristic feature of BCs: antecedent gives conditions stating when the consequent is relevant In the case of (2) it is understood that the speaker assumes the assertion that pizza is in the fridge to be relevant to the listener only in case s/he is hungry. A unified theory of ICs and BCs should account for the (in)dependence of the truth of antecedent and consequent in the two cases as well as for the observed relevance effects. WCCFL 27, 16-18 May 2008, UCLA 3 Indicative Conditionals vs. Biscuit Conditionals introduction frame setting setting –– conclusion conclusion introduction –– topics topics as as speech speech acts acts –– ICs ICs as as aboutness aboutness topics topics –– BCs BCs as as frame Biscuit conditionals (BCs) have been analysed as conditional assertions by de Rose & Grandy (1999), and as involving existential quantification over potential literal acts by Siegel (2006): IC BC GLD HTLD Aboutness Frame Setting (2) If you are hungry, (*then) there is pizza in the fridge. (3) If you are hungry, there exists an (presupposed relevant) assertion of „there is pizza in the fridge“. Problem with these accounts: they are too weak. (4) If you don‘t want to watch the movie, the gardener is the killer. (5) If the congregation is ready, I hereby declare you man and wife. Consequent speech acts have been performed unconditionally. WCCFL 27, 16-18 May 2008, UCLA 4 Two Kinds of Topic Marking Constructions introduction frame setting setting –– conclusion conclusion introduction –– topics topics as as speech speech acts acts –– ICs ICs as as aboutness aboutness topics topics –– BCs BCs as as frame IC BC GLD HTLD Aboutness Frame Setting German left dislocation (GLD): (6) Den The-ACC Pfarrer, pastor den kann keiner leiden. RP-ACC can nobody like. ‚The pastor nobody likes.‘ Hanging topic left dislocation (HTLD): (7) Der/den Pfarrer, The(-ACC) pastor keiner kann ihn nobody can him leiden. like. ‚The pastor, nobody likes him.‘ WCCFL 27, 16-18 May 2008, UCLA 5 Two Kinds of Topic Marking Constructions introduction frame setting setting –– conclusion conclusion introduction –– topics topics as as speech speech acts acts –– ICs ICs as as aboutness aboutness topics topics –– BCs BCs as as frame Frey (2004) notes that the following syntactic characteristics set apart GLD from HTLD: IC BC GLD HTLD Aboutness Frame Setting (RP) GLD requires the presence of a resumptive pronoun (preferably in Spec., CP), which is a weak d-pronoun (der, den, die, das, …). (B) GLD allows for binding into the dislocated phrase from within the clause, while HTLD does not: (8) Seineni His-ACC Vater, den father RP-ACC verehrt jederi. admires everybody. [GLD] ‚Everybody admires his father.‘ (9) *Sein(en)i Vater, jederi verehrt ihn. His(-ACC) father everybody admires him. WCCFL 27, 16-18 May 2008, UCLA [HTLD] 6 Two Kinds of Topic Marking Constructions introduction frame setting setting –– conclusion conclusion introduction –– topics topics as as speech speech acts acts –– ICs ICs as as aboutness aboutness topics topics –– BCs BCs as as frame Frey (2004) shows: GLD marks aboutness topicality, i.e. it IC BC GLD HTLD Aboutness Frame Setting establishes the entity the sentence is about. In contrast: HTLD marks frame setting topics, i.e. it establishes a frame of interpretation for which the following material is relevant. In this respect, HTLD is an instance of more general frame setting constructions: (10) As for the pastor, the marriage sermon was wonderful. Fact expressed in the matrix clause is implied to be relevant w.r.t. (questions regarding) the pastor. Note: in (10) there is no proform in the matrix clause which could pick up the entity marked as the topic. WCCFL 27, 16-18 May 2008, UCLA 7 ICs and GLD vs. BCs and HTLD introduction frame setting setting –– conclusion conclusion introduction –– topics topics as as speech speech acts acts –– ICs ICs as as aboutness aboutness topics topics –– BCs BCs as as frame IC BC GLD HTLD Aboutness Frame Setting The same syntactic characteristics set apart ICs and BCs: (RP) There is a strong syntactic similarity between ICs and correlative constructions, which ‚involve a free relative clause adjoined to the matrix clause and coindexed with a proform inside it‘ (Bhatt and Pancheva 2001). Then can therefore be regarded as a proform which relates back to the possibilities introduced by the if-clause (see e.g. also Iatridou 1994) WCCFL 27, 16-18 May 2008, UCLA 8 ICs and GLD vs. BCs and HTLD introduction frame setting setting –– conclusion conclusion introduction –– topics topics as as speech speech acts acts –– ICs ICs as as aboutness aboutness topics topics –– BCs BCs as as frame IC BC GLD HTLD Aboutness Frame Setting The same syntactic characteristics set apart ICs and BCs: (B) While binding into the if-clause is possible in the case of ICs, it is not in the case of BCs (cf. Haegeman, 2003) : jede Orchideei (11) Wenn man siei gut pflegt, dann blüht if one it well groom then blossoms every orchid mehrmals several times [IC] im Jahr. in the year ,Every orchid blossoms several times a year, if you groom it well.’ (12) *Wenn Du etwas über siei wissen willst, if you something about it to know want jede Orchideei blüht mehrmals every orchid blossoms several times WCCFL 27, 16-18 May 2008, UCLA [BC] im Jahr. in the year 9 ICs and GLD vs. BCs and HTLD introduction frame setting setting –– conclusion conclusion introduction –– topics topics as as speech speech acts acts –– ICs ICs as as aboutness aboutness topics topics –– BCs BCs as as frame IC BC Close connection between conditionals and topicality has been observed frequently (see e.g. Haiman 1978 and Bittner 2001). GLD HTLD Aboutness Frame Setting We conclude: ICs instantiate aboutness topicality BCs instantiate frame setting topicality the if-clause serves as the topic, while the matrix clause supplies comment Concerning BCs, equivalent frame setting paraphrases are possible: (2) If you are hungry, (*then) there is pizza in the fridge. (13) As for the possibility that you are hungry, there is pizza in the fridge. Semantic and pragmatic effects are completely parallel: − antecedent/frame setting topic establishes the conditions for relevance of the matrix speech act − matrix clause is asserted unconditionally WCCFL 27, 16-18 May 2008, UCLA 10 ICs and GLD vs. BCs and HTLD introduction frame setting setting –– conclusion conclusion introduction –– topics topics as as speech speech acts acts –– ICs ICs as as aboutness aboutness topics topics –– BCs BCs as as frame IC BC GLD HTLD Aboutness Frame Setting We derive the semantic and pragmatic contributions of ICs and BCs by extending the approach to aboutness topics by Endriss (to appear) to frame setting topics and combining it with the approach to indicative conditionals by Schlenker (2004), who treats if-clauses as definite descriptions of possible worlds. WCCFL 27, 16-18 May 2008, UCLA 11 Topics as Speech Acts introduction topics –– BCs BCs as as frame frame setting setting –– conclusion conclusion introduction – – topics topics as as speech speech acts acts –– ICs ICs as as aboutness aboutness topics Endriss (to appear): aboutness topics are interpreted in a separate speech act of topic establishment REFX (cf. reference act of Searle, 1969 or act of frame setting of Jacobs, 1984; cf. also Lambrecht, 1994). REFX − applies to the denotation of the topic-marked constituent and introduces a novel discourse referent X for it − is conceived as being performed before the speech act of the original utterance The comment is interpreted in the original speech act where topical discourse referent X is supplied as an argument. The two speech acts are conjoined via speech act conjunction &. Schematically for an assertion: ASSERT(COMMENT(TOPIC)) REFX(TOPIC) & ASSERT(COMMENT(X)) This approach is reminiscent of the two steps in categorical judgements WCCFL 27, 16-18 May 2008, UCLA 12 Topics as Speech Acts introduction topics –– BCs BCs as as frame frame setting setting –– conclusion conclusion introduction – – topics topics as as speech speech acts acts –– ICs ICs as as aboutness aboutness topics Endriss and Hinterwimmer (to appear): the d-pronoun in the specifier of the matrix-CP is interpreted like a relative pronoun, i.e. it triggers lambda-abstraction. In a sentence such as (6), the matrix clause is thus interpreted as shown in (14), while the entire sentence is interpreted as shown schematically in (15): (6) Den The-ACC Pfarrer, den pastor RP-ACC kann keiner leiden. can nobody like. [GLD] ‚The pastor nobody likes‘. (14) ’den kann keiner leiden÷ = λy. ¬∃z[human(z) ∧ like(z,y)] (15) REFX(ιx[pastor(x)]) & ASSERT(¬∃z[human(z) ∧ like(z,X)]) WCCFL 27, 16-18 May 2008, UCLA 13 Topics as Speech Acts introduction topics –– BCs BCs as as frame frame setting setting –– conclusion conclusion introduction – – topics topics as as speech speech acts acts –– ICs ICs as as aboutness aboutness topics Topical constituent definite / of type e: − discourse referent X in REF act can straightforwardly refer to topical constituent − no truth conditional effect Topical constituent indefinite / of generalized quantifier (GQ) type 〈〈e,t〉,t〉: − discourse referent X in REF act refers to a suitable representative that is created from the GQ, namely a minimal witness set of the topical GQ − truth conditional effect: topical GQ takes widest (possibly island-free) scope (16) Einen Song von Bob Dylan, den kennt jeder. Some-ACC song of Bob Dylan, RP-ACC knows everybody. [GLD] ‚Everybody knows some song of Bob Dylan.‘ (17) REFX(∃x[song_of_BD(x)]) & ASSERT(∀y[human(y) → know(y,X)] WCCFL 27, 16-18 May 2008, UCLA 14 Topics as Speech Acts introduction topics –– BCs BCs as as frame frame setting setting –– conclusion conclusion introduction – – topics topics as as speech speech acts acts –– ICs ICs as as aboutness aboutness topics The HTLD example in (7), is interpreted as shown in (18): (7) Der/den Pfarrer, keiner kann ihn leiden. The-(ACC) pastor nobody can him like. [HTLD] ‚The pastor nobody likes.‘ (18) REFX (ιx[pastor(x)]) & ASSERT(¬∃z[human(z) ∧ like(z, y)]) Here, the act of topic establishment has exactly the same effect. Crucial difference w.r.t. matrix clause: no resumptive proform → no lambda-abstraction → no predication w.r.t. topical discourse referent X. Proform ihn (him) in (7) is thus treated as a free variable that needs to be resolved. (Remember: No binding into BCs, see (8) and (9)) Since the topic provides the most salient discourse referent in (7), the free variable y is most likely resolved to it. WCCFL 27, 16-18 May 2008, UCLA 15 Topics as Speech Acts introduction topics –– BCs BCs as as frame frame setting setting –– conclusion conclusion introduction – – topics topics as as speech speech acts acts –– ICs ICs as as aboutness aboutness topics Hence the case of frame setting does not correspond to a simple predicative relation of topic and comment, as becomes even clearer by considering the interpretation of (10) shown in (19): (10) As for the pastor, the marriage sermon was wonderful. (19) REFX (ιx[pastor(x)]) & ASSERT(wonderful(marriage_sermon)) Since (10) does not contain any proform whatsoever, the matrix clause interpreted as an independent assertion. Only connection of topic and comment: their consecutive performance as speech acts. Here the issue of relevance comes into play: using standard Gricean assumptions, an assertion is only felicitous if it is relevant to the preceding discourse / assertion must serve to answer the question under discussion. WCCFL 27, 16-18 May 2008, UCLA 16 Topics as Speech Acts introduction topics –– BCs BCs as as frame frame setting setting –– conclusion conclusion introduction – – topics topics as as speech speech acts acts –– ICs ICs as as aboutness aboutness topics (10) As for the pastor, the marriage sermon was wonderful. (19) REFX (ιx[pastor(x)]) & ASSERT(wonderful(marriage_sermon)) In (10), the pastor is established as the topic (QUD: ‚What about the pastor?‘) The following assertion has to be relevant with respect to the QUD/has to partially answer the QUD. This is exactly the pragmatic effect we observe in the case of frame setting constructions. Note that in the case of aboutness topics, the relevance condition is trivially fullfilled, because a predication is obviously relevant to its argument: (15) REFX (ιx[pastor(x)]) & ASSERT(¬∃z[human(z) ∧ like(z, X)]) WCCFL 27, 16-18 May 2008, UCLA 17 The antecedents of ICs as aboutness topics introduction –– topics topics –– BCs introduction topics as as speech speech acts acts –– ICs ICs as as aboutness aboutness topics BCs as as frame frame setting setting –– conclusion conclusion We follow Schlenker (2004), who builds on Stalnaker (1968), in analysing the antecedents of conditionals as definite descriptions of possible worlds. if ϕ : the unique possible world which is most similar to the actual world w0 among all possible worlds where ϕ is true. The if-clause in (1) thus denotes the object in (20): the unique world which is most similar to the actual world among all posible worlds where Peter went shopping. (1) If Peter went shopping, then there is pizza in the fridge. (20) ’If Peter went shopping÷ = ιw0w[went_shopping(w)(peter)] WCCFL 27, 16-18 May 2008, UCLA [IC] 18 The antecedents of ICs as aboutness topics introduction topics –– BCs introduction –– topics topics as as speech speech acts acts –– ICs ICs as as aboutness aboutness topics BCs as as frame frame setting setting –– conclusion conclusion The proposition denoted by the consequent is then applied to the denotation of the if-clause, which gives (21): (1) If Peter went shopping, then there is pizza in the fridge. [IC] (21) pizza_is_in_the_fridge(ιw0w[went_shopping(w)(peter)]) “The unique world where Peter went shopping which is most similar to the actual world is among all the worlds where pizza is in the fridge”. Our analysis: if-clause in (1) is actually interpreted as the aboutness topic: − the act of topic establishment introduces a discourse referent X for the unique world denoted by the if-clause, and − it is then asserted that the predicate of worlds denoted by the consequent holds of X. WCCFL 27, 16-18 May 2008, UCLA 19 The antecedents of ICs as aboutness topics introduction topics –– BCs introduction –– topics topics as as speech speech acts acts –– ICs ICs as as aboutness aboutness topics BCs as as frame frame setting setting –– conclusion conclusion Proform then in Spec., CP triggers lambda-abstraction over possible worlds (parallel to d-pronouns which trigger lambda-abstraction over individuals). Hence the consequent forms a predicate that applies to the topical discourse referent (parallel to the individual aboutness topic case) and we get (22). (1) If Peter went shopping, then there is pizza in the fridge. (22) REFX(ιw0w[went_shopping(w)(peter)]) & ASSERT(pizza_is_in_the_fridge(X)) As in the case of GLD: − due to definitness/type of topic: no truth-conditional effect compared to (21) pizza_is_in_the_fridge(ιw0w[went_shopping(w)(peter)]) − relevance requirement trivially fulfilled due to predication WCCFL 27, 16-18 May 2008, UCLA 20 The antecedents of BCs as frame setting topics introduction topics –– BCs conclusion introduction –– topics topics as as speech speech acts acts –– ICs ICs as as aboutness aboutness topics BCs as as frame frame setting setting –– conclusion Crucially, BCs do not allow the proform then in the consequent (cf. Bhatt & Pancheva, 2001). As in the case of frame setting, the consequent is thus not interpreted as a predicate that applies to the topical discourse referent, but as an independent assertion. Our example (2) is therefore interpreted as shown in (23): (2) If you are hungry, (*then) there is pizza in the fridge. (23) REFX(ιw0w[hungry(w)(listener)]) & ASSERT(pizza_is_in_the_fridge(w0)) That there is pizza in the fridge is thus asserted unconditionally, which is exactly what we observed for BCs. This explains the restrictions on the occurence of then. If then is inserted into a BC it turns into an (usually pragmatically odd) IC, which is predicted in our account. WCCFL 27, 16-18 May 2008, UCLA 21 Conclusion introduction BCs as as frame frame setting setting –– binding binding effects effects –– conclusion conclusion introduction –– aboutness aboutness topics topics –– ICs ICs as as aboutness aboutness topics topics –– BCs Differences and similarities between sentences with aboutness topics and frame setting topics are parallel to the differences and similarities between ICs and BCs. Conclusion: ICs are aboutness topics BCs are frame setting topics Common aspect of all topical sentences (including ICs and BCs): (a) a speech act of topic establishment is conjoined with an assertion and (b) the assertion has to be relevant with respect to the topic Differences: (a) aboutness topics: assertion involves a predicate applied to the topic (b) frame setting topics: assertion is related to topic establishment only by relevance. WCCFL 27, 16-18 May 2008, UCLA 22 Conclusion introduction BCs as as frame frame setting setting –– binding binding effects effects –– conclusion conclusion introduction –– aboutness aboutness topics topics –– ICs ICs as as aboutness aboutness topics topics –– BCs Thank you. WCCFL 27, 16-18 May 2008, UCLA 23 Appendix: an explanation of the binding effects binding binding effects effects –– formal formal detail detail Recall that binding into the dislocated element is possible in the case of GDL and ICs, but impossible in the case of HTLDs and BCs: (8) Seineni His-ACC Vater, den father RP-ACC verehrt jederi. admires everybody. (9) *Sein(en)i Vater, jeder verehrt ihni. His(-ACC) father everybody admires him. (11) Wenn man siei gut pflegt, dann blüht jede Orchideei mehrmals If one it well grooms then blossom every orchid several times [IC] im Jahr. in the year (12) *Wenn Du etwas über siei wissen willst, jede Orchideei blüht If you something about it to know want every orchid blossoms [BC] mehrmals im Jahr. several times in the year WCCFL 27, 16-18 May 2008, UCLA 24 Appendix: an explanation of the binding effects binding binding effects effects –– formal formal detail detail Concerning GLDs, we proceed along the lines of (Ebert and Endriss, 2007) They argue that cases like (8) exemplify functional topics (8) Seinen Vateri, den verehrt jederi. His-ACC father RP-ACC admires everybody. They show that dislocated constituents with bound pronouns do not take narrow scope w.r.t. the binding matrix quantifier but functional wide scope, which can be truth-conditionally distinguished from genuine narrow scope in case of dislocated quantificational phrases. Hence the correct analysis must treat the left dislocated phrase as a function that is introduced as the aboutness topic of the sentence via the REF act. WCCFL 27, 16-18 May 2008, UCLA 25 Appendix: an explanation of the binding effects binding binding effects effects –– formal formal detail detail In case of (8) the dislocated definite denotes a function of type 〈e,e〉. The d-pronoun correspondingly is of the same type and combines with the matrix verb, which is type-shifted by application of Jacobson´s Z-operator (Jacobson, 1999), as shown for (8) in (24): (8) Seineni Vater, den verehrt jederi. His-ACC father RP-ACC admires everybody. (24) a. Z = λR〈〈e,e〉,t〉 λf〈e,e〉 λxe R(f(x))(x) b. ’Seinen Vater÷ = λy.ιx[father_of(x,y)] c. ’den verehrt jeder÷ = λf〈e,e〉.∀y[human(y) → Z(admire)(f )(y)] = λf〈e,e〉.∀y[human(y) → admire(y, f(y))] WCCFL 27, 16-18 May 2008, UCLA 26 Appendix: an explanation of the binding effects binding binding effects effects –– formal formal detail detail The sentence as a whole is accordingly interpreted as shown in (25): (8) Seineni His-ACC Vater, father den RP-ACC verehrt jederi. admires everybody. (25) REFX(λy.ιx[father_of(x,y)]) & ASSERT(∀y[human(y) → admire(y, X(y))) In case of HTLDs like (9): shifting the topical constituent to functional type does not result in binding, since the matrix clause does not denote a predicate that could be applied to this function. (9) *Sein(en)i Vater, His(-ACC) father jeder verehrt ihni. everybody admires him. (26) REFX(λy.ιx[father_of(x,y)]) & ASSERT(∀y[human(y) → admire(y, z)) (9) can thus only be interpreted in the following way: first, the function from individuals into their fathers is established as the topic and then it is asserted that everybody admires some salient individual. Since it is hard to find a context where this is coherent, the sentence is odd. WCCFL 27, 16-18 May 2008, UCLA 27 Appendix: an explanation of the binding effects binding binding effects effects –– formal formal detail detail The analysis for ICs runs entirely parallel. As the dislocated if-clause contains a pronoun it denotes a function from individuals into worlds of type 〈e,s〉. then again denotes a corresponding variable triggering lambda-abstraction and the combines with the matrix verb that is type-shifted via a adapted version Z‘ of Jacobson´s Z, as shown for (11) in (27): (11) Wenn man sie gut pflegt, dann blüht jede Orchidee mehrmals im Jahr. (27) a. Z‘ = λR〈〈e,e〉,t〉 λf〈e,s〉λxe. R(f(x))(x) b. ’Wenn man sie gut pflegt÷ = λx. ιw0w[well_groomed(w)(x)] c. ’dann blüht jede Orchidee mehrmals im Jahr÷ = λf〈e,s〉.∀y[orchid(y) → Z‘(blossom_several_times_in_year)(f )(y)] = λf〈e,s〉.∀y[orchid(y) → blossom_several_times_in_year( f (y))(y)] WCCFL 27, 16-18 May 2008, UCLA 28 Appendix: an explanation of the binding effects binding binding effects effects –– formal formal detail detail The sentence as a whole is accordingly interpreted as shown in (28): (11) Wenn man sie gut pflegt, dann blüht jede Orchidee mehrmals im Jahr. (28) REFX(λx.ιw0w[well_groomed(w)(x)]) & ASSERT(∀y[orchid(y) → blossom_several_times_in_year(X(y))(y)] In case of BCs like (12): shifting the if-clause to the functional type does not result in binding, since the matrix clause does not denote a predicate that could be applied to this function. (12) *Wenn Du etwas über sie wissen willst, jede Orchidee blüht mehrmals im Jahr. (29) REFX(λx.ιw0w[want_to_know_sth_about(w)(listener, x)]) & ASSERT(∀y[orchid(y) → blossom_several_times_in_year(w0)(y)]) (12) can thus only be interpreted in the following way: a function from individuals x into the world which is closest to the actual world where the listener wants to know something about x is established and then it is asserted that every vase is valuable in the actual world. Again, this is incoherent and hence the sentence is odd. WCCFL 27, 16-18 May 2008, UCLA 29 Appendix: formal detail binding binding effects effects – – formal formal detail detail WCCFL 27, 16-18 May 2008, UCLA 30 Appendix: formal detail binding binding effects effects – – formal formal detail detail WCCFL 27, 16-18 May 2008, UCLA 31 Appendix: formal detail binding binding effects effects – – formal formal detail detail WCCFL 27, 16-18 May 2008, UCLA 32 Appendix: formal detail binding binding effects effects – – formal formal detail detail WCCFL 27, 16-18 May 2008, UCLA 33 Appendix: formal detail binding binding effects effects – – formal formal detail detail WCCFL 27, 16-18 May 2008, UCLA 34 References Bittner, M. (2001). Topical Referents for Individuals and Possibilities. In: R. Hastings et al (eds). Proceedings from SALT XI, pp. 36-55. Bhatt, R. and R. Pancheva (2001). Conditionals. In: M. Everaert and H. van Riemsdijk (eds.), The Blackwell Companion to Syntax, vol. I, pp. 638–687. Blackwell. Ebert, C. & C. Endriss (2004): Topic Interpretation and Wide Scope Indefinites. In: Proceedings of NELS 34. Ebert, C. & C. Endriss (2007): Functional Topics. In: Proceedings of the Sinn und Bedeutung 11. Endriss, C. & S. Hinterwimmer (to appear): Indefinites as Direct and Indirect Aboutness Topics (appears in a volume on information structure edited by C. Fery and M. Zimmermann, Oxford University Press). Endriss, C. (to appear). Quantificational Topics, Studies in Linguistics and Philosophy, Springer. Frey, W. (2004). Notes on the Syntax and the Pragmatics of German Left Dislocation. In: H. Lohnstein and S. Trissler (eds.). The Syntax and Semantics of the Left Periphery, pp. 203-233. Mouton de Gruyter. Haiman, J. (1978). Conditionals are Topics. Language 54:565–589. Haegeman, L. (2003). Conditional Clauses: External and Internal Syntax. Mind & Language, Vol. 18 No. 4, pp. 317-339. Jacobs, J. (1984). Funktionale Satzperspektive und Illokutionssemantik. Linguistische Berichte 91. WCCFL 27, 16-18 May 2008, UCLA 35 References Jacobson, P.: 1999, Towards a Variable-Free Semantics, Linguistics & Philosophy 22(2), 117–184 Kratzer, A. (1986). Conditionals. Chicago Linguistics Society, 22(2): 1–15. Lambrecht, K. (1994). Information Structure and Sentence Form. Cambridge, Massachusetts: Cambridge University Press. Lewis, D. (1973). Counterfactuals and Comparative Possibility. Journal of Philosophical Logic 2, 418446. Nolan, D. (2003). Defending a Possible-Worlds Account of Indicative Conditionals. Philosophical Studies, 116: 215–269. Schlenker, P. (2004). Conditionals as Definite Descriptions, Research in Language and Computation 2:417–462. Searle, J. (1969). Speech Acts: An Essay in the Philosophy of Language. Cambridge University Press, London. Siegel, M. (2006). Biscuit Conditionals. Quantification over Potential Literal Acts. Linguistics and Philosophy 29:167–203. Iatridou, S. (1994). On the Contribution of Conditional Then. Natural Language Semantics 2:171–199. Stalnaker, R. (1968). A Theory of Conditionals. Studies in Logical Theory, American Philosophical Quarterly, Monograph: 2. Warmbrod, K. (1983). Epistemic Conditionals. Pacific Philosophical Quarterly 64: 249–265 WCCFL 27, 16-18 May 2008, UCLA 36